Bayesian Networks

 

 

 

 

A simple Bayesian network: 1 + 2 + 2 + 4 + 4:

1. 1 is the probability of getting a required parameter smoking, smoking or not smoking is known not to breathe smoke is 0.6 to 0.4 chanting so we know it is a √

2. 2 are two parameters of lung cancer when he was only related to smoking and smoking (suffering from lung cancer / not suffering from lung cancer) is a non-smoker how much time (suffering from lung cancer / not suffering from lung cancer) is how much will eventually be only two parameters number four

3. 4 is the number of parameters difficulty breathing, the probability of lung cancer and bronchitis, dyspnea xxxxxxxx00 01 10 11

The 2 * 5

 

 

Markov model is a directed acyclic graph, showing the current node only Bayesian network node and a related front. Markov is different from the network

 

 

 

 

tail to tail

p(a,b,c) = p(c)p(a|c)p(b|c)

Both sides / p (c)

And because p (ab | c) joint probability of ab = p (a, b, c) / p (c)

Therefore, to obtain p (a, b, c) / p (c) = p (a | c) p (b | c) p at c given set of conditions (a, b) = p (a) p (b) ab i.e. independent conditional independence

 

 

 

 

head to tail

p(a,b,c) = p(a)p(c|a)p(b|c)

Known p (a, b | c) = p (a, b, c) / p (c) into the p (a, b, c) = p (a) p (c | a) p (b | c )

And because p (a) p (c | a) = p (a, c)

To = p (a, c) p (b | c) / p (c) it = p (a | c) p (b | c) = p (a, b | c) at c given too case ab blocked independent

 

 

 head to head

p(a,b,c) = p(a)p(b)p(c|a,b)

Equate to the left and right sides of the case make the integral of c p (a, b) = p (a) p (b) and that is independent of c, ab independent

Similar to a family in the absence of parents with children is independent of both parents have contacted